
% Author: Kazuma Takakura
% Date: 17 May 2025
% Time: 10:47:58

\begin{table}[h!]\footnotesize
  \centering
  \caption{Internal Validity in the Presence of Attrition}
\label{tab:ghanem_attrition}
\scalebox{1}{
\begin{threeparttable}
\begin{tabular}{lcccccc}\toprule
   & \multicolumn{4}{c}{Mean Baseline Outcome by Group} & \multicolumn{2}{c}{p-value} \\ 
 \cmidrule(lr){2-5} \cmidrule(lr){6-7}  Dependent Variable & TR & CR & TA & CA & IVal-R & IVal-P   \\\midrule\midrule
 DT score^{a} & 0.132 & 0.207 & -0.057 & -0.048 & 0.845 & 0.035 \\ 
                               & [ 0.972 ] & [ 0.919 ] & [ 1.010 ] &  [ 1.013 ] & &    \\ 
 RSES^{a} & 0.015 & -0.081 & -0.022 & 0.037 & 0.551 & 0.712 \\ 
                               & [ 0.981 ] & [ 1.045 ] & [ 0.929 ] &  [ 1.071 ] & &    \\ 
 CPCS^{a} & 0.109 & -0.241 & 0.062 & -0.042 & 0.010 & 0.023 \\ 
                               & [ 0.971 ] & [ 1.124 ] & [ 0.912 ] &  [ 1.059 ] & &    \\ 
 Female & 0.607 & 0.612 & 0.563 & 0.589 & 0.755 & 0.710 \\ 
                               & [ 0.490 ] & [ 0.490 ] & [ 0.497 ] &  [ 0.493 ] & &    \\ 
 Grade 3 & 0.614 & 0.663 & 0.592 & 0.534 & 0.205 & 0.073 \\ 
                               & [ 0.489 ] & [ 0.475 ] & [ 0.492 ] &  [ 0.499 ] & &    \\ 
 \\ 
\midrule
\end{tabular}
\begin{tablenotes}
\item (a) Variables are standardized using the average and variance of the whole baseline sample in the March 2016 survey. 
\item (b) Standard deviations are reported in square brackets.
\item (c) The mean baseline outcomes correspond to the groups of treatment respondents (TR), control respondents (CR), treatment attritors (TA), and control attritors (CA).
\end{tablenotes}
\end{threeparttable}
}
\label{tab:addlabel}%
\end{table}




